[net.news] New! Improved! Line-Eater bug!!

throopw@dg_rtp.UUCP (Wayne Throop) (10/18/86)

The Line-Eater lives!!!  Well, not *THE* line eater, mind you.  Just a
distant relative.  This one likes to munch on articles from the end
towards the beginning.  The basis for this one's snacking habits are
also not known.  Nor its unnatural habitat.  I noticed this mangled
article in mod.comp-soc

    Path: dg_rtp!rti-sel!mcnc!philabs!cmcl2!yale!
          husc6!rutgers!sri-spam!sri-unix!hplabs!hplabsc!taylor
    Newsgroups: mod.comp-soc
    Subject: Virtual Communities and so on.
    Message-ID: <719@hplabsc.UUCP>
    Lines: 127

The article really seems to have about 80 lines (actually
79-and-a-partial), and seemed to be truncated at 3669 characters
(excluding header), and 4158 with header.  It is possible that at some
intermediate node the article size was exactly 4096 at the cut-off
point, which might be interesting.

In order to have another manifestation of the bug to examine, I looked
around more-or-less at random and found

    Path: dg_rtp!rti-sel!mcnc!philabs!cmcl2!seismo!think!nike!
          cit-vax!ll-xn!mit-amt!mit-eddie!mit-vax!oaf
    Newsgroups: talk.politics.misc
    Subject: All the Muslims I know???
    Message-ID: <894@mit-vax.UUCP>
    Lines: 67

This one seemed to really have around 29 lines.  Without header it
seemed to be 1615 characters, and with header about 2157.  *This* one
might well have been 2048 characters at some intermediate site.

Any comment from the system admins of the systems shared on the two
paths?  Recent new software?  Hardware problems?  Hmmmmm?

--
Software is under a constant tension.  Being symbolic it is arbitrarily
perfectible; but also it is arbitrarily changeable.
                                --- Alan J. Perlis
-- 
Wayne Throop      <the-known-world>!mcnc!rti-sel!dg_rtp!throopw

grr@cbmvax.cbm.UUCP (George Robbins) (10/20/86)

In article <653@dg_rtp.UUCP> throopw@dg_rtp.UUCP (Wayne Throop) writes:
>The Line-Eater lives!!!  Well, not *THE* line eater, mind you.  Just a
>distant relative.  This one likes to munch on articles from the end
>towards the beginning.  The basis for this one's snacking habits are
>also not known.  Nor its unnatural habitat.  I noticed this mangled
>article in mod.comp-soc
>
>
>The article really seems to have about 80 lines (actually
>79-and-a-partial), and seemed to be truncated at 3669 characters
>(excluding header), and 4158 with header.  It is possible that at some
>intermediate node the article size was exactly 4096 at the cut-off
>point, which might be interesting.

Any reason not to assume it's the boring (but painful) 'oops, excuse me
while my spool filesystem fills up' syndrome?  Sure would be nice if this
new 2.11 stuff purges articles that had errors on writing to disk, rather
than leaving them lay about for re-transmission...
-- 
George Robbins - now working for,	uucp: {ihnp4|seismo|caip}!cbmvax!grr
but no way officially representing	arpa: cbmvax!grr@seismo.css.GOV
Commodore, Engineering Department	fone: 215-431-9255 (only by moonlite)

gds@sri-spam.istc.sri.com (The lost Bostonian) (10/20/86)

I checked and the article <719@hplabsc.UUCP> arrived here intact.
/usr/spool doesn't fill up here anymore, so I doubt it got munched for
that reason.

--gregbo

wex@milano.UUCP (10/20/86)

In article <653@dg_rtp.UUCP>, throopw@dg_rtp.UUCP (Wayne Throop) writes:
> The Line-Eater lives!!!  Well, not *THE* line eater, mind you.  Just a
> distant relative.  This one likes to munch on articles from the end
> towards the beginning.

The problem Wayne describes has also been noticed here (Austin, Texas).
It happens a lot in talk.philosophy.misc, too.

-- 
Alan Wexelblat
ARPA: WEX@MCC.ARPA or WEX@MCC.COM
UUCP: {seismo, harvard, gatech, pyramid, &c.}!ut-sally!im4u!milano!wex

"I'm a peeping-tom techie with x-ray eyes"

desj@brahms (David desJardins) (10/21/86)

In article <653@dg_rtp.UUCP> throopw@dg_rtp.UUCP writes:
>The article really seems to have about 80 lines (actually
>79-and-a-partial), and seemed to be truncated at 3669 characters
>(excluding header), and 4158 with header.  It is possible that at some
>intermediate node the article size was exactly 4096 at the cut-off
>point, which might be interesting.

   Presumably a disk filled up on one of the intermediate nodes, and so
the article was truncated when it was written into the spool directory.
Then the truncated version was retransmitted and found its way to you.

   -- David desJardins

shaddock@rti-sel.UUCP (Mike Shaddock) (10/22/86)

In article <653@dg_rtp.UUCP> throopw@dg_rtp.UUCP (Wayne Throop) writes:
[I've trimmed the article down -- Mike Shaddock]
>The Line-Eater lives!!!  Well, not *THE* line eater, mind you.  Just a
>distant relative.  This one likes to munch on articles from the end
>
>    Path: dg_rtp!rti-sel!mcnc!philabs!cmcl2!yale!
>          husc6!rutgers!sri-spam!sri-unix!hplabs!hplabsc!taylor
>    Newsgroups: mod.comp-soc
>    Subject: Virtual Communities and so on.
>    Message-ID: <719@hplabsc.UUCP>
>
>    Path: dg_rtp!rti-sel!mcnc!philabs!cmcl2!seismo!think!nike!
>          cit-vax!ll-xn!mit-amt!mit-eddie!mit-vax!oaf
>    Newsgroups: talk.politics.misc
>    Subject: All the Muslims I know???
>    Message-ID: <894@mit-vax.UUCP>
>-- 
>Wayne Throop      <the-known-world>!mcnc!rti-sel!dg_rtp!throopw

I checked both of these articles at rti-sel and mcnc.  They had arrived
truncated at both places.
-- 
Mike Shaddock	{decvax,seismo,ihnp4,akgua,philabs}!mcnc!rti-sel!shaddock

"You're in a twisty maze of sendmail rules, all obscure."

throopw@dg_rtp.UUCP (Wayne Throop) (10/23/86)

> grr@cbmvax.cbm.UUCP (George Robbins)
>> throopw@dg_rtp.UUCP (Wayne Throop)

>>The Line-Eater lives!!!  Well, not *THE* line eater, mind you.  Just a
>>distant relative.
> Any reason not to assume it's the boring (but painful) 'oops, excuse me
> while my spool filesystem fills up' syndrome?

It might well be.  But there have been so many of them lately, and over
such a span of time that it seemed unlikely that a single disk space
crunch episode could explain the problem.  Of course, if many of the
truncated articles passed through a system with a *chronic* disk space
crunch in its recent history, then "The problem is sol-ved."

> Sure would be nice if this
> new 2.11 stuff purges articles that had errors on writing to disk, rather
> than leaving them lay about for re-transmission...

Say "Amen!", somebody.

--
Usenet:  all the news that's fit to garble.
-- 
Wayne Throop      <the-known-world>!mcnc!rti-sel!dg_rtp!throopw